Zero-Results Rate

Zero Results · No-Results Rate · Null Results Rate · Zero-Hit Rate

Definition

The zero-results rate is a site-search metric: the share of search queries for which the system found no results, out of all queries over a period.

How it is calculated

The zero-results rate is a core quality metric for site search:

Zero-results rate = queries with no results / all search queries × 100%

Example: in a week shoppers ran 40,000 queries and 1,200 of them returned an empty page, so the zero-results rate is 3%. Count all queries, not distinct phrasings: otherwise a hundred rare typos outweigh the single frequent query on which the store loses the most.

Why search finds nothing

Cause	Example query	What to do
Typo	“sneekers”	Fuzzy matching
Wrong keyboard layout	“zoga mat” typed on German QWERTZ	Layout conversion
Regional spelling	“jewellery” when the catalogue says “jewelry”	Spelling-variant dictionary
Word form	“knives” when the catalogue says “knife”	Stemming or lemmatisation
Synonym	“hoodie”, while the catalogue says “hooded sweatshirt”	Synonym list, semantic search
Query too narrow	“blue sneakers size 13”	Relax the conditions, show the nearest match
Product not stocked	“skis” in an electronics store	Pass the query to buying

Zero results as demand data

Split empty queries into two streams. “Search did not understand” is technical debt: every cause you fix closes a whole group of phrasings at once. “Not stocked” is a ready-made list of what shoppers wanted to buy from you specifically and could not. That is input for category managers and buying.

What to show instead of an empty page

An empty page is a dead end: the shopper can only rephrase or leave, and that shows up in your bounce rate. Instead, show:

the corrected query, with a note that results are for the corrected version;
related categories;
products closest to the query in meaning;
a recommendation widget with bestsellers or personalized picks.

Checklist

Track the metric weekly, separately for desktop and mobile.
Review the top 50 empty queries by hand and tag the cause.
After every search release and every large catalogue upload, run a control set of queries.
Make sure autosuggest never leads to an empty results page.
Read the metric together with results CTR: 500 irrelevant products technically count as zero empty queries, but they do not help the shopper.

Frequently asked questions

What zero-results rate counts as normal?

There is no universal benchmark: the figure depends on assortment breadth, the share of branded queries and how repeats are counted. It is more useful to compare the store with itself — week on week and after every search change. Check the 100 most frequent queries separately: none of them should return nothing.

How do I tell a search failure from a missing product?

Check the query by hand: if the product is in the catalogue but was not found, it is a failure (typo, synonym, word form, keyboard layout). If the product is not stocked, it is an assortment request. A simple spreadsheet with a “cause” column works well; once a month, hand the second type of queries to category managers.

What should the shopper see instead of an empty page?

A corrected query with a note that it was corrected, related categories, products close in meaning and a recommendation block — bestsellers or personalized picks. Be honest that there is no exact match: a silent substitution looks like a search bug.

Should the zero-results rate be counted by query or by session?

Both versions are useful. By query, you see which phrasings break search. By session — the share of search sessions with at least one empty search — you see how many shoppers hit a dead end. For the full picture, add the share of those sessions that ended with the shopper leaving the site.

Gravity Field

Fewer zero results with AI Search

Gravity Field AI Search runs on vector semantic search, so instead of an empty page the shopper gets products close to the query in meaning.

✦A lower zero-results rate through meaning-based matches

✦Vector semantic search across the product catalogue

✦Typo tolerance — resilient to misspellings and spelling variants

✦Autosuggest driven by the intent behind the query

Request a demo →